Are you seriously ready for AI and ML in Incident Automation? Part 2

Processes and Procedures to Automate Incidents
Getting clean, precise data - is probably the easiest step on your journey toward automating incidents.

In the last blog, “Can you trust your data to lead you to AIOps?” we covered the first step in the process of arriving at automation, which is building your ITOps on a foundation of clean, accurate data. However, what we neglected to mention is that step – getting clean, precise data – is probably the easiest step on your journey toward automating incidents. Don’t get me wrong: data is a significant milestone, but it’s just one marker toward your goal. 

Step #2: Device Discovery

Brian Amaro

The second step necessitates using the right tools to discover the devices, applications, third-party interfaces and additional elements that make up your network.

Discovering every device, in real-time, is crucial – yet difficult to do – in a modern IT ecosystem where technologies are rapidly spinning up and down and producing data that needs to be collected, standardized, and properly stored. Only when those pieces have been discovered and properly indexed will you be able to follow up with the necessary checks and quality controls, locking down the inputs to ensure they remain reliable.

Unfortunately, the work doesn’t end there. 

Step 3: Processes and Procedures

The third step is taking that clean data then creating and implementing policies and procedures that enable the organizational agility needed to drive automation. Said differently, once you get your hands on the data, you need an implementation plan – and here’s where most companies stumble.

According to a recent Gartner survey, 91%  of organizations haven’t achieved “transformational” use of their data, meaning that–despite being a priority investment–data and analytics are not making much of a difference in operations. The problem is not a technology issue, but one of confronting and overcoming human nature.

When you engage fully in the transformational journey from ITOps to AIOps and ask people to act based on the knowledge that comes from precise data, things get uncomfortable. The data can uncover design flaws, knowledge gaps, and expose the accumulated legacy of suboptimal decisions. And that’s a huge risk for anyone who adapted to those conditions.

What to expect when initiating processes and procedures

In their 2013 Analysis of the Factors Affecting Resistance to Change in Management Accounting Systems, Drs. Rodrigo Angonese and Carlos Eduardo Facin Lavarda identified seven factors contributing to organizational resistance to change:

  1. Institutional power; 2. Ontological insecurity; 3. Trust; 4. Inertia; 5. Lack of knowledge; 6. Acceptance of routines; and, 7. Decoupling or loose coupling of required new practices.

The changes required to turn the data into productive, actionable intelligence touch every one of these factors. But to get there means establishing programs that address the resistance to change and processes that keep your data, and operations, honest.

Gaining broad support for your organization’s journey to AIOps means sending a clear message that participation in the processes leading to data-driven transformation is not a pretext for punishing past behavior. Instead, it’s a chance to start afresh and apply institutional knowledge to improve associated business outcomes; an opportunity to reform old, inefficient routines with efficient automation; a chance to eliminate the frustrating and engage in the fulfilling.

Articulating a vision for what the journey means for the future of your organization and instilling enthusiasm for that journey at all levels will, with follow-through, build trust and overcome the inertia associated when doubt is present. Executives, line managers and IT staff alike must see that the changes are instrumental and not a threat to their current status.

“Learn from the past and apply those lessons to the future”

The path toward automating incidents is realized when it’s based on policies that address old, inefficient systems, recognizing and fixing the flaws present in “the way we do things around here” mentality.

Old processes are why we had unreliable data, shadow IT and a devil-may-care attitude toward onboarding new devices, applications, and out-of-scope cloud instances. New policies should set a proper baseline for data, ensure the accuracy of configuration management databases, maintain compliance with administrative rules and third-party terms of service and, ultimately, keep your data honest.

Wrap-up:

With the right policies, that honest data will provide the means to implement AIOps and offer clear metrics for measuring success. Faster incident reporting and earlier, more accurate trouble reporting; lower mean-time to resolution; lower mean-time between incidents; speedier transactioning. Honest data is the difference between today’s reactive posture and tomorrow’s proactive posture. It means automating the human factor out of the drudgery and reallocating those high-value resources to more high-value activities.

Accenture’s 2018 report, The Future Belongs to Intelligent Operations, found that nearly 80% of companies fear “disruption” by more digitally savvy enterprises. That strongly suggests an opportunity to be among the 20% leading the way. You get there with honest data, deliberate processes, and logical policies as your stepping stones from IT Ops to AIOps and truly efficient automation.